Decades of research at the tertiary level have produced the finding that the use of different teaching methods makes no difference to student learning when end-ofcourse examination performance is accepted as the criterion of such learning. This conclusion has been based on a comparison between the average mark gained by groups of subjects taught in different ways. The present article takes the position that the use of such mean scores serves to conceal individual variability, and hypothesizes that it is possible to identify personality characteristics likely to predispose a student to perform better under one teaching approach rather than another. A programmatic series of studies using two approaches to personality measurement, purpose-concealed and face-valid, and two instructional approaches, lecturing and independent study, is described culminating in the conclusion that the student likely to perform better under the latter teaching method sees himself as being more tense and anxious than the student likely to perform better under the former method.